Opportunistic infections and AIDS malignancies early after initiating combination antiretroviral therapy in high-income countries The HIV-CAUSAL Collaboration Background: There is little information on the incidence of AIDS-defining events which have been reported in the literature to be associated with immune reconstitution inflammatory syndrome (IRIS) after combined antiretroviral therapy (cART) initiation. These events include tuberculosis, mycobacterium avium complex (MAC), cytomegalovirus (CMV) retinitis, progressive multifocal leukoencephalopathy (PML), herpes simplex virus (HSV), Kaposi sarcoma, non-Hodgkin lymphoma (NHL), cryptococcosis and candidiasis. Methods: We identified individuals in the HIV-CAUSAL Collaboration, which includes data from six European countries and the US, who were HIV-positive between 1996 and 2013, antiretroviral therapy naive, aged at least 18 years, had CD4þ cell count and HIVRNA measurements and had been AIDS-free for at least 1 month between those measurements and the start of follow-up. For each AIDS-defining event, we estimated the hazard ratio for no cART versus less than 3 and at least 3 months since cART initiation, adjusting for time-varying CD4þ cell count and HIV-RNA via inverse probability weighting. Results: Out of 96 562 eligible individuals (78% men) with median (interquantile range) follow-up of 31 [13,65] months, 55 144 initiated cART. The number of cases varied between 898 for tuberculosis and 113 for PML. Compared with non-cART initiation, the hazard ratio (95% confidence intervals) up to 3 months after cART initiation were 1.21 (0.90–1.63) for tuberculosis, 2.61 (1.05–6.49) for MAC, 1.17 (0.34–4.08) for CMV retinitis, 1.18 (0.62–2.26) for PML, 1.21 (0.83–1.75) for HSV, 1.18 (0.87–1.58) for Kaposi sarcoma , 1.56 (0.82–2.95) for NHL, 1.11 (0.56–2.18) for cryptococcosis and 0.77 (0.40–1.49) for candidiasis. Conclusion: With the potential exception of mycobacterial infections, unmasking IRIS does not appear to be a common complication of cART initiation in high-income ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins countries.
AIDS 2014, 28:2461–2473 Keywords: HIV, immune reconstitution inflammatory syndrome, incidence, inverse probability weighting, unmasking
Introduction Combined antiretroviral therapy (cART) has dramatically reduced morbidity and mortality associated with HIV infection [1–4]. cART restores the immune response against opportunistic infections, but some patients experience an inflammatory reaction within weeks or months after cART initiation [5,6]. This immune reconstitution inflammatory syndrome (IRIS),
whose pathogenesis is not fully elucidated, can result in clinical worsening of existing opportunistic infections after commencing cART (paradoxical IRIS) or in the appearance soon after cART initiation of a new and previously unrecognised opportunistic infections (unmasking IRIS) [7]. IRIS, may be associated with significant morbidity, is a diagnostic challenge and complicates clinical management [7].
Correspondence to Dr Sara Lodi, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA. E-mail: [email protected] Received: 23 May 2014; revised: 13 August 2014; accepted: 13 August 2014. DOI:10.1097/QAD.0000000000000456
ISSN 0269-9370 Q 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
2461
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
2462
AIDS
2014, Vol 28 No 16
IRIS has been described in patients with opportunistic infections and AIDS malignancies caused by infections [8–17] as well as in patients with non-infectious conditions, such as rheumatoid arthritis and sarcoidosis, although with a different immunopathogenesis [18,19]. Whereas there is a solid body of literature documenting reactions associated with immune restoration for mycobacterial infections both in the HIV-negative [20] and the HIV-positive population [21–24], our knowledge of IRIS for other conditions is mainly based upon case studies of patients on cART. Most of these describe cases of paradoxical phenomena and only a few studies have reported on unmasking IRIS. Further, because case definitions have not been implemented in large observational databases, it has been problematic to estimate its magnitude. The HIV-CAUSAL Collaboration has recently reported an increase in tuberculosis incidence shortly after cART initiation which was particularly marked in patients with CD4þ cell counts below 50 cells/ml, a pattern strongly suggestive of unmasking IRIS [25]. This pattern was not seen for Pneumocystis jirovecii pneumonia. Here we update and extend our study to the effect of cART on AIDSdefining events suggested to be associated with IRIS in the literature: tuberculosis, mycobacterium avium complex (MAC), cytomegalovirus (CMV) retinitis, progressive multifocal leukoencephalopathy (PML), herpes simplex virus (HSV), Kaposi sarcoma, non-Hodgkin lymphoma (NHL), cryptococcosis and candidiasis. For each of these AIDS-defining events, we explore whether changes in incidence after cART initiation are compatible with unmasking IRIS.
Methods Study population We used data from the HIV-CAUSAL collaboration, which include HIV-positive individuals from prospective cohorts in six European countries and the United States [25]. All cohorts are based on routinely collected data in clinical practice within settings with universal access to care. Initiation of cARTwas defined as the date on which an individual initiated treatment with at least two nucleoside reverse transcriptase inhibitors and either one or more protease inhibitors, one non-nucleoside reverse transcriptase inhibitors, one entry/fusion inhibitor or one integrase inhibitor. Analyses included individuals who were HIV-positive between 1996 and 2013 aged 18 years or more, and had a CD4þ cell count and an HIV-RNA measurement within 6 months of each other while antiretroviral therapy (ART) naive. Individuals’ follow-up started at baseline, defined as the date when all the inclusion criteria were met, and ended at outcome diagnosis, death, 12 months
after the most recent laboratory measurement or cohortspecific administrative censoring, whichever occurred earlier. To prevent the misclassification of undiagnosed prevalent opportunistic infections and AIDS malignancies as incident cases, and thus, minimise the inclusion of cases of paradoxical IRIS, our analyses excluded HIV-positive individuals who were not AIDS-free during the baseline month.
Outcomes We considered as primary outcomes all AIDS-events previously suggested to be associated with IRIS. We included tuberculosis, CMV retinitis, cryptococcosis, PML and Kaposi sarcoma because these were the most common AIDS-defining events in a systematic review of IRIS in observational studies [11]. We included MAC because its association with IRIS was observed soon after antiretroviral therapy was introduced [21]. We included NHL because rare manifestations of IRIS have been reported [10,26]. We included candidiasis and HSV (often not considered in association with IRIS) because a large cohort study of cART initiators in the United States [12] reported them as the most common IRIS-related events. The diagnostic criteria for the AIDS-defining events [27] were those routinely used in clinical practice in each of the participating countries. Information on use of prophylaxis drugs for these conditions is not collected by the HIV-CAUSAL Collaboration because prophylaxis for these conditions is not widely implemented in most of the participating cohorts. For each outcome, our working definition for unmasking IRIS was a newly diagnosed and non-previously detected AIDS-defining event in the first 3 months after starting cART.
Statistical methods All analyses were conducted separately for each outcome. We computed incidence rates as number of cases per 1000 person-years and estimated the hazard ratio of each outcome for (i) cART versus no cART and (ii) no cART versus less than 3 and at least 3 months since cART initiation. We then estimated the cumulative incidence up to 3 months after cART initiation [28]. To estimate the hazard ratios, we used a pooled logistic model for risk of the outcome at month mþ1 that included a time-varying indicator for ever use of cART through month m, month of follow-up m (restricted cubic splines with five knots) and the following baseline covariates: CD4þ cell count (
AIDS 2014, 28:2461–2473 Keywords: HIV, immune reconstitution inflammatory syndrome, incidence, inverse probability weighting, unmasking
Introduction Combined antiretroviral therapy (cART) has dramatically reduced morbidity and mortality associated with HIV infection [1–4]. cART restores the immune response against opportunistic infections, but some patients experience an inflammatory reaction within weeks or months after cART initiation [5,6]. This immune reconstitution inflammatory syndrome (IRIS),
whose pathogenesis is not fully elucidated, can result in clinical worsening of existing opportunistic infections after commencing cART (paradoxical IRIS) or in the appearance soon after cART initiation of a new and previously unrecognised opportunistic infections (unmasking IRIS) [7]. IRIS, may be associated with significant morbidity, is a diagnostic challenge and complicates clinical management [7].
Correspondence to Dr Sara Lodi, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA. E-mail: [email protected] Received: 23 May 2014; revised: 13 August 2014; accepted: 13 August 2014. DOI:10.1097/QAD.0000000000000456
ISSN 0269-9370 Q 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
2461
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
2462
AIDS
2014, Vol 28 No 16
IRIS has been described in patients with opportunistic infections and AIDS malignancies caused by infections [8–17] as well as in patients with non-infectious conditions, such as rheumatoid arthritis and sarcoidosis, although with a different immunopathogenesis [18,19]. Whereas there is a solid body of literature documenting reactions associated with immune restoration for mycobacterial infections both in the HIV-negative [20] and the HIV-positive population [21–24], our knowledge of IRIS for other conditions is mainly based upon case studies of patients on cART. Most of these describe cases of paradoxical phenomena and only a few studies have reported on unmasking IRIS. Further, because case definitions have not been implemented in large observational databases, it has been problematic to estimate its magnitude. The HIV-CAUSAL Collaboration has recently reported an increase in tuberculosis incidence shortly after cART initiation which was particularly marked in patients with CD4þ cell counts below 50 cells/ml, a pattern strongly suggestive of unmasking IRIS [25]. This pattern was not seen for Pneumocystis jirovecii pneumonia. Here we update and extend our study to the effect of cART on AIDSdefining events suggested to be associated with IRIS in the literature: tuberculosis, mycobacterium avium complex (MAC), cytomegalovirus (CMV) retinitis, progressive multifocal leukoencephalopathy (PML), herpes simplex virus (HSV), Kaposi sarcoma, non-Hodgkin lymphoma (NHL), cryptococcosis and candidiasis. For each of these AIDS-defining events, we explore whether changes in incidence after cART initiation are compatible with unmasking IRIS.
Methods Study population We used data from the HIV-CAUSAL collaboration, which include HIV-positive individuals from prospective cohorts in six European countries and the United States [25]. All cohorts are based on routinely collected data in clinical practice within settings with universal access to care. Initiation of cARTwas defined as the date on which an individual initiated treatment with at least two nucleoside reverse transcriptase inhibitors and either one or more protease inhibitors, one non-nucleoside reverse transcriptase inhibitors, one entry/fusion inhibitor or one integrase inhibitor. Analyses included individuals who were HIV-positive between 1996 and 2013 aged 18 years or more, and had a CD4þ cell count and an HIV-RNA measurement within 6 months of each other while antiretroviral therapy (ART) naive. Individuals’ follow-up started at baseline, defined as the date when all the inclusion criteria were met, and ended at outcome diagnosis, death, 12 months
after the most recent laboratory measurement or cohortspecific administrative censoring, whichever occurred earlier. To prevent the misclassification of undiagnosed prevalent opportunistic infections and AIDS malignancies as incident cases, and thus, minimise the inclusion of cases of paradoxical IRIS, our analyses excluded HIV-positive individuals who were not AIDS-free during the baseline month.
Outcomes We considered as primary outcomes all AIDS-events previously suggested to be associated with IRIS. We included tuberculosis, CMV retinitis, cryptococcosis, PML and Kaposi sarcoma because these were the most common AIDS-defining events in a systematic review of IRIS in observational studies [11]. We included MAC because its association with IRIS was observed soon after antiretroviral therapy was introduced [21]. We included NHL because rare manifestations of IRIS have been reported [10,26]. We included candidiasis and HSV (often not considered in association with IRIS) because a large cohort study of cART initiators in the United States [12] reported them as the most common IRIS-related events. The diagnostic criteria for the AIDS-defining events [27] were those routinely used in clinical practice in each of the participating countries. Information on use of prophylaxis drugs for these conditions is not collected by the HIV-CAUSAL Collaboration because prophylaxis for these conditions is not widely implemented in most of the participating cohorts. For each outcome, our working definition for unmasking IRIS was a newly diagnosed and non-previously detected AIDS-defining event in the first 3 months after starting cART.
Statistical methods All analyses were conducted separately for each outcome. We computed incidence rates as number of cases per 1000 person-years and estimated the hazard ratio of each outcome for (i) cART versus no cART and (ii) no cART versus less than 3 and at least 3 months since cART initiation. We then estimated the cumulative incidence up to 3 months after cART initiation [28]. To estimate the hazard ratios, we used a pooled logistic model for risk of the outcome at month mþ1 that included a time-varying indicator for ever use of cART through month m, month of follow-up m (restricted cubic splines with five knots) and the following baseline covariates: CD4þ cell count (
